From owner-freebsd-questions@FreeBSD.ORG Mon Mar 7 10:39:56 2005 Return-Path: Delivered-To: freebsd-questions@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id C81BC16A4CE for ; Mon, 7 Mar 2005 10:39:56 +0000 (GMT) Received: from rwcrmhc12.comcast.net (rwcrmhc12.comcast.net [216.148.227.85]) by mx1.FreeBSD.org (Postfix) with ESMTP id 9B3CC43D5E for ; Mon, 7 Mar 2005 10:39:56 +0000 (GMT) (envelope-from emccoy@haystacks.org) Received: from [127.0.0.1] (c-24-98-109-41.atl.client2.attbi.com[24.98.109.41]) by comcast.net (rwcrmhc12) with ESMTP id <2005030710395401400jilfge>; Mon, 7 Mar 2005 10:39:55 +0000 Message-ID: <422C2F74.5020407@haystacks.org> Date: Mon, 07 Mar 2005 05:39:48 -0500 From: Eric McCoy User-Agent: Mozilla Thunderbird 0.8 (Windows/20040913) X-Accept-Language: en-us, en MIME-Version: 1.0 To: dave References: <000501c522b7$23baa040$0800a8c0@satellite> In-Reply-To: <000501c522b7$23baa040$0800a8c0@satellite> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit cc: FreeBSD Questions Subject: Re: apache exits on signal 4 X-BeenThere: freebsd-questions@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: User questions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 07 Mar 2005 10:39:56 -0000 dave wrote: > Hello, > Fbsd 5.3-RELEASE, apache 2.53, php4 installed. Everything was working > than i got a report of site's being down. I investigated and learned that > apache was not starting. I tried a restart, which did not produce an error, > however checking for an http process ID didn't show anything. I then checked > /var/log/messages and /var/log/httpd-access.log and found that the apache > startup was dying on a signal 4. Googling on this did not tell me what it > was or more importantly how to fix it. Signal 4 is SIGILL, illegal instruction. (man 3 signal) If you've been able to run the same httpd+modules successfully this is probably indicative of a hardware failure. Usually memory, but it could be your CPU or motherboard or a variety of other things. If this is your hardware and you can afford the downtime, put memtest86 on a boot disk or CD and run it overnight to see if it spots anything. You could also try running Apache through truss and seeing if it gets the error in the same place every time. You could recompile Apache with debugging code, but that will probably make this error go away (by changing the code just enough to obscure the underlying problem). If you are running in a jail, get your hosting company to look at the host system's dmesg/logs since they may see warnings about memory parity errors which might otherwise not be passed to your jail. Linux folks also say that compiling the kernel is a good "stress test." Usually if there's a memory or other hardware problem it will cause gcc to die. I expect that doing a make world would have similar effects. You could try that if all else fails.